Challenge

Managing over 25K claims– and processing around 10K– a year isn’t easy. The process involves many tedious tasks, such as categorising requests, responding to emails and extracting data from documents. This situation created not only more workload for employees but also a delayed response time for airline customers.

Solution

Aviclaim chose Qubiz due to our track record of creating process automation solutions based on AI, that allow employees to switch from manual tasks to high-value activities.

24/7 first-level support through a chatbot

Aviclaim needed a faster way to offer first-level support for known issues and to guide customers through the flight claim process.

We developed a chatbot for the Aviclaim websites and those of their affiliate partners using natural language processing (NLP) to understand user conversation intent.

To create and train models for language understanding, we used the Microsoft Bot Builder framework and LUIS, they’re a machine-learning service.

Speeding up claim payment prediction

Previously, Aviclaim employees had to take each email and analyse which claims have a higher chance of compensation.

To speed up this process, we created the claim payment prediction feature to automatically predict which claims will be successful.

How does it work? Using ChatGPT for predictive analysis. We worked with a dataset from their system and enriched it with other features derived from the existing ones or created by us. First, the system uses the airline response classification (such as “claim acceptance”) then it uses both Aviclaim-customer and airline-customer email exchanges to predict whether or not the claim request will be accepted.

Eliminating tedious tasks with Airline response classification

Aviclaim employees had to manually categorise claims, a highly tedious and time-wasting process.

Instead, using a ChatGPT integration, scanning and interpreting incoming airline responses is much more efficient. The agent now classifies email responses into categories such as “claim acceptance” or “claim rejection.”

Quick customer inquiry responses

Timely responses=happy customers. To ensure this, we worked on an email bot that can understand and respond to customer inquiries.

The ChatGPT-based bot uses the client’s database, Aviclaim emails and airline conversations to offer an overview of everything related to a claim. In this way, customers get quick and accurate answers to their emails, regarding the status of their claim, delays, documents, etc.

Simplifying the document handling process

A large percent of Aviclaim employees’ time is spent verifying and extracting information from documents, such as passports, and responding to customer emails.

We are developing a ChatGPT solution to handle customer inquiries, classify airline responses and handle legal documents (missing or incomplete information).

For example, if a claim has missing information or documents (ID, plane tickets), the agent will automatically send an email asking customers for their information.

Additionally, the machine learning models are tested for accuracy and reliability, as they will continuously learn new customer questions, minimising errors over time.

Looking to the future

As part of our continuous development and support process, we are currently working on adding more AI features and training models for increased efficiency and accuracy.

Results

With our help, Aviclaim can now:

• Benefit from cutting-edge AI/ML solutions with 24/7 support
• Make better decisions regarding which claims to accept
• Handle claims and documentation efficiently and more accurately with much less manual processing
• Offer support and guide customers through the process
• Benefit from a painless claims process leading to more satisfied customers

Testimonial